In silico quantification of cell proportions from mixed-cell transcriptomics data (deconvolution) requires a reference expression matrix, called basis matrix. We hypothesize that matrices created using only healthy samples from a single microarray platform would introduce biological and technical biases in deconvolution. We show presence of such biases in two existing matrices, IRIS and LM22, irrespective of deconvolution method. Here, we present immunoStates, a basis matrix built using 6160 samples with different disease states across 42 microarray platforms. We find that immunoStates significantly reduces biological and technical biases. Importantly, we find that different methods have virtually no or minimal effect once the basis matrix is chosen. We further show that cellular proportion estimates using immunoStates are consistently more correlated with measured proportions than IRIS and LM22, across all methods. Our results demonstrate the need and importance of incorporating biological and technical heterogeneity in a basis matrix for achieving consistently high accuracy.

A major contributor to the scientific reproducibility crisis has been that the results from homogeneous, single-center studies do not generalize to heterogeneous, real world populations. Multi-cohort gene expression analysis has helped to increase reproducibility by aggregating data from diverse populations into a single analysis. To make the multi-cohort analysis process more feasible, we have assembled an analysis pipeline which implements rigorously studied meta-analysis best practices. We have compiled and made publicly available the results of our own multi-cohort gene expression analysis of 103 diseases, spanning 615 studies and 36,915 samples, through a novel and interactive web application. As a result, we have made both the process of and the results from multi-cohort gene expression analysis more approachable for non-technical users.

A major contributor to the scientific reproducibility crisis has been that the results from homogeneous, single-center studies do not generalize to heterogeneous, real world populations. Multi-cohort gene expression analysis has helped to increase reproducibility by aggregating data from diverse populations into a single analysis. To make the multi-cohort analysis process more feasible, we have assembled an analysis pipeline which implements rigorously studied meta-analysis best practices. We have compiled and made publicly available the results of our own multi-cohort gene expression analysis of 103 diseases, spanning 615 studies and 36,915 samples, through a novel and interactive web application. As a result, we have made both the process of and the results from multi-cohort gene expression analysis more approachable for non-technical users.

BACKGROUND: This study assessed the use of active surveillance in men with low-risk prostate cancer and evaluated institutional factors associated with the receipt of active surveillance. METHODS: A retrospective, hospital-based cohort of 115,208 men with lowrisk prostate cancer diagnosed between 2010 and 2014 was used. Multivariate and mixed effects models were used to examine variation and factors associated with active surveillance. RESULTS: During the study period, the use of active surveillance increased from 6.8% in 2010 to 19.9% in 2014 (estimated annual percentage change, 128.8%; 95% confidence interval [CI], 1 19.6% to 1 38.7%; P 5.002). The adjusted probability of active-surveillance receipt by institution was highly variable. Compared with patients treated at comprehensive community cancer centers, patients treated at community cancer programs (odds ratio [OR], 2.00; 95% CI, 1.50-2.67; P <.001) and academic institutions (OR, 2.47; 95%, CI, 1.81-3.37; P <.001) had higher odds of receiving active surveillance. Compared with patients treated at very low-volume facilities, patients treated at very high-volume facilities had higher odds of receiving active surveillance (OR, 3.57; 95% CI, 1.94-6.55; P <.001). Patient and hospital characteristics accounted for 60.2% of the overall variation, whereas the treating institution accounted for 91.5% of the unexplained variability. CONCLUSIONS: Within this hospital-based cohort, the use of active surveillance for low-risk prostate cancer increased significantly over time. Significant variation was found in the use of active surveillance. Most of the variation was attributable to facility-related factors such as the facility type, facility volume, and institution. Policies to achieve consistent and higher rates of active surveillance, when appropriate, should be a priority of professional societies and patient advocacy groups. Cancer 2018;124:55-64.